CN101420406A - Disturbance eliminating and balancing method between carrier waves, equipment and OFDM receiver - Google Patents

Abstract

A method and device for eliminating and equalizing inter-carrier interference, and an OFDM receiver. A method for eliminating inter-carrier interference ICI of the kth subcarrier of Orthogonal Frequency Division Multiplexing OFDM of a received signal (k is a natural number greater than 1 and less than or equal to n, and n is the number of subcarriers), comprising: passing decision feedback The equalized DFE method eliminates the intercarrier interference element produced by at least one subcarrier in the intercarrier interference element of the k subcarrier; eliminates the intercarrier interference element produced by at least one other subcarrier by the linear minimum mean square error equalization LMMSE method Another element of intercarrier interference caused by subcarriers.

Description

Inter-carrier interference elimination and equalization methods, equipment and OFDM receiver

The application requires in the priority of the 10-2007-0107556 korean patent application of submission on October 25th, 2007, and it openly is contained in this, for reference.

Technical field

The disclosure relates to the method that inter-carrier interference (ICI) is eliminated or compensated, and more particularly, relates to the two step ICI removing methods that have low complex degree and superperformance in time varying channel (that is mobile channel).

Background technology

OFDM (OFDM) is the strong technology that is used for transmitting effectively by channel data.A plurality of subcarrier frequencies (subcarrier) in this technology use channel width are with the transmission data.With in order to divide and to isolate subcarrier spectrum and can waste traditional frequency division multiplexing (FDM) of local channel bandwidth and compare, these subcarriers are arranged reaching optimum bandwidth efficient, thereby avoid inter-carrier interference (ICI).

On the contrary, though the frequency spectrum of OFDM subcarrier is significantly overlapping in OFDM channel width scope, OFDM allows decomposition and the recovery of modulated to the information of each subcarrier.

Use ofdm signal to cause tolerance limit, the effective spectrum of multidiameter delay expansion and frequency selective attenuation are used, simplify the balanced and good interference characteristic of subchannel by channel transmission data.

Yet OFDM requires the point-device Frequency Synchronization between receiver and the transmitter; Because frequency departure, subcarrier is quadrature no longer, causes the ICI between the subcarrier or crosstalks.

Usually cause frequency shift (FS) by unmatched transmitter and receiver oscillator or owing to move the Doppler frequency shift that causes.Although can be only by the receiver compensating for doppler frequency shifts, worse off when adding multipath makes it more be difficult to proofread and correct because reflection appears in various frequency shift (FS) meeting.

To increase this influence even worse along with speed, and this influence be the use of restriction OFDM in hot-short (as, the mobile TV in automobile, bus, train etc. uses) element.

Use OFDM to have several various criterions as the mobile TV of broadcast interface, as, the hand-held digital video broadcast of popularizing fast in Europe (DVB-H), cross the FLO that the single forward link of medium (FLO, Forward Link Only) system provides mobile TV service in the digital media broadcast (DVB) of Korea S's exploitation and at Americanologist.

In order to improve the performance of OFDM receiver, the pilot signal of some subcarrier portability measured channel conditions in some OFDM symbols (that is, gain of the equalizer of each subcarrier and phase shift).Pilot signal and training symbol also can be used for time synchronized (to avoid intersymbol interference, ISI) and Frequency Synchronization (with the inter-carrier interference of avoiding causing, ICI) by Doppler frequency shift.

Fig. 1 does not have the functional block diagram that ICI eliminates traditional OFDM receiver of element.

With reference to Fig. 1, traditional OFDM receiver 10 can comprise: down-converter 1, to descending conversion by the signal of changing on the OFDM transmitter (not shown); Analog/digital converter (ADC) 2 will be a digital signal from the conversion of signals of down-converter 1 output; Protect (GI) remover 3 at interval, removing OFDM transmitter (not shown) from the signal of ADC2 output is the protection interval of channel estimating input; Fast Fourier transform (FFT) unit 4 carries out fast fourier transform to the signal from 3 outputs of GI remover; Channel estimator 6 is estimated channel frequency response based on the FFT signal of FFT unit 4 conversions; Equalizer 5, based on the channel frequency response from channel estimator output, the signal that FFT is changed carries out equilibrium; Decoder 7 is decoded to the signal from the equilibrium of equalizer 5 output.

As shown in Figure 1, equalizer 5 can be used for static environment; Yet in mobile environment, the performance of equalizer 5 significantly reduces, because it does not have the circuit that inter-carrier interference ICI eliminates.

A kind of method that reduces the ICI influence is used the OFDM symbol of shorter duration.Yet because the channel delay expansion, this method produces little effect and can not be used for single frequency network.

The ICI removing method can roughly be divided into two kinds.A kind of is linear minimum mean-squared error LMMSE method, and another kind is a judgement-feedback error DFE method.

The LMMSE method is close to perfect in the advantage aspect the elimination ICI; Yet it requires high complexity.On the other hand, estimate the ICI element and carry out the method for eliminating the ICI element of estimating from the signal that receives repeatedly that by preliminary ruling or hard decision the DFE method has the superperformance of relatively low complexity.Yet, in the DFE method, when the wrong large percentage of preliminary judgement or hard decision, because carry out the error propagation effect that there is the performance that reduces whole OFDM receiver in DFE repeatedly.Therefore, a kind of method and apparatus that ICI eliminates that is used for that overcomes the shortcoming of above-mentioned ICI removing method and utilize each advantage most effectively of expectation.

Summary of the invention

Exemplary embodiment of the present tries hard to provide the effective I CI removing method of the suitable low complex degree of a kind of relative pinpoint accuracy of the LMMSE of having method and DFE method.

According to an exemplary embodiment of the present, (k is greater than 1 and be less than or equal to n for the inter-carrier interference ICI removing method of the k subcarrier of a kind of OFDM of received signal, n is the quantity of subcarrier), comprising: eliminate the inter-carrier interference element that produces by at least one subcarrier in the inter-carrier interference element of k subcarrier by decision feedback equalization DFE method; Eliminate in the inter-carrier interference element of k subcarrier another inter-carrier interference element that produces by another subcarrier at least by the balanced LMMSE method of linear minimum mean-squared error.

The step of eliminating by the LMMSE method can comprise: eliminate by k subcarrier to the (the inter-carrier interference element that the subcarrier of k ± L) produces (L is more than or equal to 1 and less than the natural number of n/2).

The step of eliminating by the DFE method can comprise: eliminate by (k ± 1) subcarrier to the (inter-carrier interference element that the subcarrier of k ± Q) produces (Q is more than or equal to 2 and less than the natural number of n/2).

The step of eliminating by the DFE method can comprise: eliminate by the (k ± (L+1)) subcarrier to the (inter-carrier interference element that the subcarrier of k ± Q) produces.L can be the natural number less than Q.LMMSE can comprise (L-order) LMMSE L time.

According to an exemplary embodiment of the present, a kind of equalization methods comprises: based on the channel frequency response from channel estimating apparatus output, the signal from the output of fast fourier transform FFT unit is carried out equilibrium; Eliminate the inter-carrier interference element that produces by at least one subcarrier in the inter-carrier interference element of k subcarrier of balanced signal by decision feedback equalization DFE method; Eliminate in the inter-carrier interference element of k subcarrier another inter-carrier interference element that produces by another subcarrier at least by the balanced LMMSE method of linear minimum mean-squared error.

The step of eliminating by the LMMSE method can comprise: eliminate by k subcarrier to the (the inter-carrier interference element that the subcarrier of k ± L) produces.

The step of eliminating by the DFE method can comprise: eliminate by (k ± 1) subcarrier to the (inter-carrier interference element that the subcarrier of k ± Q) produces.The step of eliminating by the DFE method can comprise: eliminate by the (k ± (L+1)) subcarrier to the (inter-carrier interference element that the subcarrier of k ± Q) produces.

According to an exemplary embodiment of the present, a kind of balancing equipment comprises: the internal balance device, based on the channel frequency response from channel estimating apparatus output, carry out equilibrium to the signal from the output of FFT unit; The DFE compensator is eliminated the ICI element of the signal of the output of equalizer internally by the DFE method; The LMMSE equalizer is eliminated from another ICI element of the signal of DFE compensator output by the LMMSE method.

The internal balance device can comprise: divider, will export divided by channel frequency response from the signal of FFT unit output.

The DFE compensator can comprise: decision device, make preliminary ruling about the signal of equalizer output internally; Multiplier will multiply each other with the channel slope vector of exporting from channel estimating apparatus from the signal of decision device output; The DFE filter carries out filtering to the inter-carrier interference element that is produced by at least one subcarrier by decision feedback equalization DFE method from the inter-carrier interference element of the k subcarrier of the signal of multiplier output; Subtracter will deduct from the signal of DFE filter output from the signal of FFT unit output.

The LMMSE equalizer can be eliminated from the ICI element of the k subcarrier of the signal of DFE compensator output another ICI element that is produced by another subcarrier at least by the balanced LMMSE method of linear minimum mean-squared error.The LMMSE equalizer can be eliminated by k subcarrier to the (the ICI element that the subcarrier of k ± L) produces.The DFE filter is eliminated by (k ± 1) subcarrier to the (ICI element that the subcarrier of k ± Q) produces.

The DFE filter can be eliminated by the (k ± (L+1)) subcarrier to the (ICI element that the subcarrier of k ± Q) produces.The LMMSE equalizer can be L LMMSE equalizer.L can be the natural number less than Q.

According to an exemplary embodiment of the present, a kind of OFDM receiver comprises: equalizer; And decoder, receive from the signal of equalizer output, and the signal that receives is decoded.

Description of drawings

By the detailed description below in conjunction with accompanying drawing, it is clear that exemplary embodiment of the present invention will become, wherein:

Fig. 1 is the functional block diagram of traditional OFDM receiver;

Fig. 2 shows the ICI power division between the adjacent sub-carriers be used to explain the ICI equalization methods according to an exemplary embodiment of the present invention;

Fig. 3 schematically shows the channel matrix of L LMMSE of ICI removing method according to an exemplary embodiment of the present invention;

Fig. 4 explains the flow chart of two step ICI removing methods according to an exemplary embodiment of the present invention;

Fig. 5 illustrates balancing equipment according to an exemplary embodiment of the present invention;

Fig. 6 illustrates the functional block diagram of OFDM receiver according to an exemplary embodiment of the present invention;

Fig. 7 and Fig. 8 have shown the simulation result that ICI removing method is according to an exemplary embodiment of the present invention compared with conventional method.

Embodiment

Below, with reference to the accompanying drawings exemplary embodiment of the present is described in detail, wherein, identical label is represented same parts all the time.To can represent by following equation from the OFDM Sn that the OFDM transmitter sends.(equation 1)

$s_n=\underset{l=0}{\overset{N-1}{\mathrm{\&Sigma;}}}{S}_l{e}^{\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A200810174934E00171.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;nl}/N}n=\mathrm{0,1},...N-1$

In equation 1, S _lCan be the quadrature amplitude modulation qam symbol in the frequency domain, N can refer to the size of discrete Fourier transform (DFT) DFT.

When receiving by time-variant multipath channel, can represent by following equation with the ofdm signal that is sent out.(equation 2)

$r_n=\underset{m=0}{\overset{N-1}{\mathrm{\&Sigma;}}}{h}_m\left(n\right)s_{n-m}+w_n,n=\mathrm{0,1},...N-1$

In equation 2, h _m(n) channel impulse response of demonstration time n.

Ofdm signal shown in the equation 2 can be handled through fixing conversion or analogue-to-digital conversion in advance.It also can or remove the protection processing of GI at interval of inserting through combination in the OFDM transmitter.

Then, signal is handled through DFT, for example, and fast fourier transform FFT, and can represent by following equation.(equation 3)

R＝Fr＝ΛS+W

In equation 3, F shows the DFT matrix, and represents R=[R respectively ₀... R _N-1] ^T, r=[r ₀... r _N-1] ^T, S=[S ₀... S _N-1] ^TIn addition, w=[W ₀... W _N-1] ^TShow the additive white Gaussian noise AWGN in the frequency domain, A shows channel matrix.The element of channel matrix can be expressed as [Λ] _{D, k}=λ (d-k, k).

Here, (equation 4)

$\mathrm{\&lambda;}(d,k)=\frac{1}{N}\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}\underset{m=0}{\overset{N-1}{\mathrm{\&Sigma;}}}{h}_m\left(n\right)e^{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A200810174934E00171.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;}(\mathrm{mk}+\mathrm{dn})/N}$

For example, λ (0, :), λ (+1, :) and λ (1 :) show respectively channel matrix leading diagonal vector, channel matrix quickly diagonal vector (lower sub-diagonal vector) and channel matrix first on sub-diagonal vector (upper sub-diagonal vector).

Can use first derivative that the element of channel matrix is similar to.Also can the coefficient of channel matrix be similar to, the element of channel matrix can be expressed as follows.

(equation 5)

λ(O，k)＝H _k，k＝1，N

(equation 6)

λ(q，k)＝ξ _qdH _k，k＝1，N

(equation 7)

$\mathrm{\&lambda;}(-q,k)={\mathrm{\&xi;}}_q^{*}d{H}_k,k=\stackrel{\&OverBar;}{1,N}$

Here, (equation 8)

${\mathrm{\&xi;}}_q=\frac{1}{2[{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A200810174934E00171.png"\; state="\{\{state\}\}">e</figure-callout>}}^j-1]}$

In addition, dH _kThe approximation that shows channel slope vector.

The LMMSE equalizer can be represented with following equation

(equation 9)

$\hat{S}={\mathrm{\&Lambda;}}^H{({\mathrm{\&Lambda;\&Lambda;}}^H+\frac{1}{{\mathrm{\&sigma;}}^2}{I}_N)}^{-1}R$

Here, σ ²It is the variance of AWGN.

Shown in equation 9, need N * N inverse of a matrix matrix operation in the LMMSE equalizer, in this case, complexity becomes N ³The order of magnitude.In using LMMSE equalizer process, so high complexity may be fatal problem, because (for example, DVB-T or DVB-H) N is approximately 8000 under the situation in actual ofdm system.

In order to reduce this high complexity, can use the approximation method of LMMSE.

This method uses the banded diagonal of channel matrix approximate, promptly, to except the leading diagonal vector, be set to 0 method by the element of all channel matrixes of (that is the sub-diagonal vector of top and following L of leading diagonal vector (L sub-diagonal vector)).

This method can be described as the LMMSE method L time.In order to calculate when carrying out equilibrium, L time the LMMSE method is only considered the ICI element that the L subcarrier (that is the 2L subcarrier of both direction) based on the subcarrier that closes on each direction produces.Because the ICI element is subjected to the distance affects between the subcarrier, so in L LMMSE, only consider to close on subcarrier.

Fig. 2 shows the ICI power division between the adjacent sub-carriers in the ICI method according to an exemplary embodiment of the present invention.With reference to Fig. 2, ICI can be based on the energy leakage between the subcarrier in the time varying channel, and this energy leakage may be subjected to the influence of the distance between the subcarrier pro rata.

For example, the subcarrier that the ICI of k subcarrier is had the greatest impact may be near (k-1) subcarrier and (k+1) subcarrier of k subcarrier.On the contrary, along with subcarrier is far away more from the k subcarrier, its influence to the ICI of k subcarrier is more little.

Therefore, the ICI method can only consider to close on the influence of the 2L subcarrier of k subcarrier when carry out calculating according to an exemplary embodiment of the present invention.Therefore, the approximated channel matrix among L LMMSE can be set to 0 except all elements the 2L+1 diagonal vector.

Fig. 3 schematically shows the shape of the channel matrix of L LMMSE of ICI removing method according to an exemplary embodiment of the present invention.With reference to Fig. 3, when initial LMMSE was N * N channel matrix, the channel matrix of L LMMSE can comprise and the sub-diagonal vector of last L of the corresponding leading diagonal vector of k subcarrier, leading diagonal vector and the sub-diagonal vector of following L of leading diagonal vector.All the other elements of channel matrix all can be set to 0.

Therefore, L LMMSE can only consider that the predetermined ICI element (and not considering the ICI element that all the other subcarriers produce) that closes on the subcarrier generation reduces complexity by carrying out equilibrium.Parameter L can be adjudicated and is trading off between complexity and the performance.Usually, can L be set to number less than 3.

Compare with the equalizer that uses in the static channel (for example, as the equalizer of not considering the ICI element of Fig. 1), above-mentioned L LMMSE method presents outstanding performance.When L was set to 1 or 2, L LMMSE method presented the performance and relative low complexity of improvement, yet this possibly can't reach the performance of full LMMSE method.

Do not consider that ICI element based on the subcarrier except the subcarrier that closes on the k subcarrier causes the performance that descends.The ICI element of the ICI removing method subcarrier generation that can use above-mentioned DFE method to eliminate in L LMMSE, not consider according to an exemplary embodiment of the present invention.

On the other hand, when full LMMSE equalizer is represented with equation 9, use L LMMSE equalizer of approximated channel matrix to represent with following equation.

(equation 10)

$\begin{array}{cc}{\hat{S}}_k=f_k^H{R}_k& k=\mathrm{0,1},...,N\end{array}$

Here, illustrate

(equation 11)

$f_k={({\mathrm{\&sigma;}}^2{I}_{2L}+{\mathrm{\&Lambda;}}_k{\mathrm{\&Lambda;}}_k^H)}^{-1}{\mathrm{\&lambda;}}_k$

R _k＝[R _k-L...R _k+L] ^T

(equation 12)

λ _k＝[λ _k-L，k…λ _k+L，k] ^T。

In addition, can be expressed as λ _{K, k}=H _k, λ _{K+l, k}=ξ _lDH _k, ${\mathrm{\&lambda;}}_{k-l,k}={\mathrm{\&xi;}}_l^{*}{\mathrm{dH}}_k.$

In other words, when L was 1, channel matrix was

${\mathrm{\&Lambda;}}_k=\left[\begin{array}{ccccc}{\mathrm{\&lambda;}}_{k-1,k-2}& {\mathrm{\&lambda;}}_{k-1,k-1}& {\mathrm{\&lambda;}}_{k-1,k}& 0& 0\\ 0& {\mathrm{\&lambda;}}_{k,k-1}& {\mathrm{\&lambda;}}_{k,k}& {\mathrm{\&lambda;}}_{k,k+1}& 0\\ 0& 0& {\mathrm{\&lambda;}}_{k+1,k}& {\mathrm{\&lambda;}}_{k+1,k+1}& {\mathrm{\&lambda;}}_{k+1,k+2}\end{array}\right],$

When L was 2, channel matrix was

${\mathrm{\&Lambda;}}_k=\left[\begin{array}{ccccccccc}{\mathrm{\&lambda;}}_{k-2,k-4}& {\mathrm{\&lambda;}}_{k-2,k-3}& {\mathrm{\&lambda;}}_{k-2,k-2}& {\mathrm{\&lambda;}}_{k-2,k-1}& {\mathrm{\&lambda;}}_{k-2,k}& 0& 0& 0& 0\\ 0& {\mathrm{\&lambda;}}_{k-1,k-3}& {\mathrm{\&lambda;}}_{k-1,k-2}& {\mathrm{\&lambda;}}_{k-1,k-1}& {\mathrm{\&lambda;}}_{k-1,k}& {\mathrm{\&lambda;}}_{k-1,k+1}& 0& 0& 0\\ 0& 0& {\mathrm{\&lambda;}}_{k,k-2}& {\mathrm{\&lambda;}}_{k,k-1}& {\mathrm{\&lambda;}}_{k,k}& {\mathrm{\&lambda;}}_{k,k+1}& {\mathrm{\&lambda;}}_{k,k+2}& 0& 0\\ 0& 0& 0& {\mathrm{\&lambda;}}_{k+1,k-1}& {\mathrm{\&lambda;}}_{k+1,k}& {\mathrm{\&lambda;}}_{k+1,k+1}& {\mathrm{\&lambda;}}_{k+1,k+2}& {\mathrm{\&lambda;}}_{k+1,k+3}& 0\\ 0& 0& 0& 0& {\mathrm{\&lambda;}}_{k+2,k}& {\mathrm{\&lambda;}}_{k+2,k+1}& {\mathrm{\&lambda;}}_{k+2,k+2}& {\mathrm{\&lambda;}}_{k+2,k+3}& {\mathrm{\&lambda;}}_{k+2,k+4}\end{array}\right].$

From equation 10 to 12 as can be known, need carry out 2L * 2L inverse matrix computing, to carry out L LMMSE method to each subcarrier of ofdm signal.Yet, shown in channel matrix, many elements are set to 0, and can use the previous result calculated of matrix channel in calculating and the corresponding channel matrix process of k subcarrier, for example, with (k-1) operation result to the corresponding channel matrix of (k-L) subcarrier.

Therefore, the actual complex degree can reach O ((1+2L) ²N).If the numerical value of L is little, then compare with full LMMSE method, L time the LMMSE method can significantly reduce complexity.Simultaneously, can eliminate in above-mentioned L the LMMSE method ICI element that subcarrier that being used to of not considering reduce complexity produces by the DFE method.

The result of the equilibrium of DFE method use standard OFDM restituted signal is with reconstruct ICI element.The DFE equalizer can be represented with following equation.

(equation 13)

$R_k^{\&prime;}=R_k-\underset{q=-Q}{\overset{Q}{\mathrm{\&Sigma;}}}{\hat{S}}_{k+q}{\mathrm{dH}}_{k+q}{\mathrm{\&xi;}}_q$

Here,

Be presented at signal by the demodulation of standard OFDM demodulation method by the result of the judgement after the equilibrium.

Right side second term shows the ICI element of k subcarrier.Therefore,

Show the signal of having eliminated the ICI element, can be right by a traditional tap equalizers Carry out equilibrium.

Shown in equation 13, along with Q increases, the DFE method is linear to be increased.Therefore, in actual DFE equalizer, Q is set up between 10 and 20.That is, available low-down complexity is estimated based on the ICI element that closes on 10 to 20 subcarriers.Yet as mentioned above, the DFE method can have the error propagation effect, this error propagation effect can cause the time become decreased performance in the multi-path environment.

The error propagation effect is subjected to percentile influence of initial mistake by the conventional equalizer generation.Promptly, when the initial wrong percentage that is produced by conventional equalizer is low relatively, eliminate the ICI element from the signal that receives and to cause performance improvement (that is, the wrong percentile minimizing that ICI causes), because the ICI element and the difference between the actual ICI element of reproducing are little.In addition, by carrying out such decision-feedback method repeatedly, and along with wrong percentage reduces, it is similar to actual ICI element that the ICI element of reproduction becomes gradually.Finally, can further reduce wrong percentage, cause performance to improve.

Yet when the initial wrong percentage that is produced by conventional equalizer was high relatively, this made that the ICI element that reproduces is incorrect, eliminates incorrect ICI element from the signal that receives, and percentile increase leads to errors.By based on the wrong percentage that increases enforcement of judgment execute a judgement feedback method repeatedly, the performance of system descends gradually.

As mentioned above, error propagation effect (disadvantage in the DFE pattern) be subjected to initial wrong percentile influence bigger.Therefore, have than higher complexity but have the LMMSE method (especially L LMMSE) of perfect performance can be used for to the influence of ICI element high relatively close on subcarrier.

On the other hand, low complex degree DFE method is used to eliminate the ICI element that subcarrier far away produces.Because each subcarrier far away causes a small amount of ICI (see figure 2), thus incorrect declare will never appreciable impact reconstruct ICI, therefore, the error propagation effect significantly reduces.

Promptly, in two step ICI removing methods according to the embodiment of the invention, in order to eliminate the ICI element of k subcarrier shown in Figure 2, the DFE method is used for than closing on the 2L subcarrier more away from the subcarrier of k subcarrier, and LMMSE method (especially L LMMSE method) is used to close on the 2L subcarrier.

Fig. 4 shows the flow chart of explaining two step ICI removing methods according to an exemplary embodiment of the present invention, and Fig. 5 shows balancing equipment according to an exemplary embodiment of the present invention.With reference to Fig. 4 and Fig. 5, at step S100, equalizer 100 of the present invention can receive from the signal R of FFT unit (not shown) output _k, from the channel frequency response H of channel estimating apparatus (not shown) output _kWith channel slope vector dH _kBalancing equipment comprises internal balance device 110, DFE compensator 120 and LMMSE equalizer 130.

At step S110, internal balance device 110 is based on carrying out from the signal R of FFT unit (not shown) output from the channel frequency response of channel estimating apparatus (not shown) output _kEquilibrium.Can use a traditional tap equalizers.One tap equalizers can be passed through R _kDivided by channel frequency response H _kCarry out equilibrium.

DFE compensator 120 can comprise decision device 121, second arithmetic facility 123, DFE filter 125 and the 3rd arithmetic facility 127.DFE compensator 120 can be eliminated the inter-carrier interference element of signal of the equilibrium of 110 outputs of equalizer internally by the DFE method.

Decision device 121 can comprise forward error correction FEC circuit (not shown).

At step S120, second arithmetic facility 123, DFE filter 125 and the 3rd arithmetic facility 127 can calculate and eliminate as Q and under the ICI element away from least one subcarrier of k subcarrier.Here, as Q and under the subcarrier away from the k subcarrier be meant the subcarrier of index between K-Q and K+Q.

Now, explained how DFE compensator 120 according to an exemplary embodiment of the present invention eliminates the ICI element of subcarrier, described subcarrier is far away and be no more than Q (that is, have from (k ± (L+1)) subcarrier to the (K ± Q) index of subcarrier) more away from the k subcarrier than L.

As shown in Figure 2, for eliminate subcarrier away from the k subcarrier (as, ± (k+L+1, k+L+2 ... k+Q) (Q herein〉L)) the ICI element, DFE compensator 120 is carried out the described processing of following equatioies.

(equation 14)

$R_k^{\&prime;}=R_k-\underset{q=L+1}{\overset{Q}{\mathrm{\&Sigma;}}}{\hat{S}}_{k+q}{\mathrm{dH}}_{k+q}{\mathrm{\&xi;}}_q-\underset{q=-Q}{\overset{-(L+1)}{\mathrm{\&Sigma;}}}{\hat{S}}_{k+q}{\mathrm{dH}}_{k+q}{\mathrm{\&xi;}}_q$

As mentioned above, most of energy of ICI element produces by closing on subcarrier, and therefore, when making improperly when closing on the judgement of subcarrier, the error propagation effect is more serious.

Therefore, the DFE method that equation 14 is described can not influence the error propagation effect, because it does not carry out the DFE method to nearest 2L subcarrier.

Yet, owing to the most influential ICI element that closes on the subcarrier generation is not eliminated, so the DFE method that is difficult to only use equation 14 to describe is expected optimum performance.Second arithmetic facility 123 that comprises in the DFE compensator 120, DFE filter 125 and the 3rd arithmetic facility 127 can be used for the right side of calculation equation 14.

Second arithmetic facility 123 can be with the signal and the channel slope vector dH that exports from the channel estimating apparatus (not shown) from decision device 121 outputs _kMultiply each other.But the ICI element that 125 filtering of DFE filter are produced by subcarrier, described subcarrier are far away and be no more than Q (Q is more than or equal to 2 and less than the natural number of n/2, n is the quantity of subcarrier) more away from the signal from 123 outputs of second arithmetic facility than L.

For this reason, DFE filter 125 also can receive the channel matrix coefficient ξ of the predetermined operation equipment (not shown) output that also comprises from balancing equipment 120 _Q... ξ _L+1Predetermined operation equipment (not shown) also can be with other channel matrix coefficient ξ _L... ξ ₁Outputing to LMMSE equalizer 130, the three arithmetic facilities 127 can deduct from the signal of DFE filter 125 outputs from the signal of FFT unit (not shown) output.

Therefore, after the ICI of DFE method eliminates processing as mentioned above, can carry out the ICI elimination processing of LMMSE method (especially L LMMSE) at step S130.

LMMSE equalizer 130 can be expressed as signal R with equation 14 for the situation of L LMMSE equalizer _kBe inserted into the equation of the equation 10 of L LMMSE method of expression.

That is, by with equation 14 $R_k^{\&prime;}={\left[\begin{array}{ccc}{R}_{k-L}^{\&prime;}& ...& R_{k+L}^{\&prime;}\end{array}\right]}^T$ Be inserted into the R of equation 10 _k=[R _K-L... R _K+L] ^TAnd obtain.

Therefore, can to eliminate from the k subcarrier from the signal of DFE compensator 120 output be interference element between the subcarrier that produces of L or subcarrier still less (L is more than or equal to 1 and less than the natural number of n/2) to LMMSE equalizer 130.

At step S140, LMMSE equalizer 130 can balanced signal outputs to the decoder (not shown) by LMMSE method (especially L LMMSE method) elimination ICI element.Compare with traditional decision-feedback method, two step ICI removing methods can reduce the error propagation effect greatly according to an exemplary embodiment of the present invention, and compare with only using Q (Q-order) LMMSE method, two step ICI removing methods can be eliminated the ICI that subcarrier far away produces by the DFE method does not have the index of complexity to increase.

Fig. 6 illustrates the functional block diagram of OFDM receiver according to an exemplary embodiment of the present invention.With reference to Fig. 6, OFDM receiver 200 can comprise aforesaid two step balancing equipment 100 and decoders 210 according to an exemplary embodiment of the present invention.Decoder 210 can receive from the signal of balancing equipment 100 outputs and to the signal that receives and decode.Decoder 210 can be embodied as the fec decoder device.OFDM receiver 200 also can comprise aforesaid down-converter, ADC, GI remover, FFT unit and/or channel estimating apparatus.

Fig. 7 and Fig. 8 have shown the simulation result that ICI removing method is according to an exemplary embodiment of the present invention compared with conventional method.With reference to Fig. 7, compare L LMMSE method decreased performance with the ICI removing method with DFE method according to an exemplary embodiment of the present invention.This is because L is not even as big as eliminating the ICI element.

In addition, traditional DFE method is carried out iteration twice, so its complexity is to the ICI removing method is similar according to an exemplary embodiment of the present invention.In this case, the ICI removing method has and the similar performance of traditional DFE method according to an exemplary embodiment of the present invention.On the other hand, as shown in Figure 8, when Doppler frequency is high, that is, when fast moving, because the error propagation effect, traditional DFE method performance seriously descends.

Compare with the LMMSE method with traditional DFE method, ICI removing method performance improves according to an exemplary embodiment of the present invention, because it eliminates the ICI element of the subcarrier generation far away relatively of low complex degree when reducing the error propagation effect.

Though shown and described exemplary embodiment of the present invention, it should be appreciated by those skilled in the art that under the situation that does not break away from the scope of the present disclosure and spirit, can carry out various modifications, interpolation and replacement.

Claims (20)

1. A method for eliminating intercarrier interference of the kth subcarrier of OFDM receiving signals, k is a natural number greater than 1 and less than or equal to n, and n is the number of subcarriers, including: Eliminating the intercarrier interference element generated by at least one subcarrier among the intercarrier interference elements of the kth subcarrier by a decision feedback equalization method; Another intercarrier interference element generated by at least another subcarrier among the intercarrier interference elements of the kth subcarrier is eliminated by a linear minimum mean square error equalization method. 2. The method according to claim 1, wherein the step of canceling by the linear minimum mean square error equalization method comprises: canceling the intercarrier interference element generated by the kth subcarrier to the (k±L)th subcarrier, L is a natural number greater than or equal to 1 and less than n/2. 3. The method according to claim 1, wherein the step of eliminating by decision feedback equalization method comprises: eliminating the intercarrier interference element generated by the kth subcarrier to the (k±Q)th subcarrier, Q is greater than or A natural number equal to 2 and less than n/2. 4. The method as claimed in claim 2, wherein the step of canceling by decision feedback equalization method comprises: canceling the carrier generated by the (k±(L+1))th subcarrier to the (k±Q)th subcarrier interfering elements, Q is a natural number greater than or equal to 2 and less than n/2. 5. The method of claim 4, wherein L is a natural number smaller than Q. 6. The method of claim 1, wherein the linear minimum mean square error equalization is L-degree linear minimum mean square error equalization. 7. An equalization method, comprising: equalizing the signal output from the fast Fourier transform unit based on the channel frequency response output from the channel estimation device; Eliminating the intercarrier interference element generated by at least one subcarrier among the intercarrier interference elements of the kth subcarrier of the equalized signal by a decision feedback equalization method; Eliminating another intercarrier interference element produced by at least another subcarrier among the intercarrier interference elements of the kth subcarrier by a linear minimum mean square error equalization method, K is a natural number greater than 1 and less than or equal to n, where n is the number of subcarriers. 8. The method according to claim 7, wherein the step of canceling by the linear minimum mean square error equalization method comprises: canceling the intercarrier interference element generated by the kth subcarrier to the (k±L)th subcarrier, L is a natural number greater than or equal to 1 and less than n/2. 9. The method as claimed in claim 7, wherein the step of canceling by decision feedback equalization method comprises: canceling the intercarrier interference element generated by the kth subcarrier to the (k±Q)th subcarrier, Q is greater than or A natural number equal to 2 and less than n/2. 10. The method according to claim 8, wherein the step of canceling by decision feedback equalization method comprises: canceling the carriers generated by the (k±(L+1))th subcarrier to the (k±Q)th subcarrier interfering element, Q is a natural number greater than L and greater than or equal to 2 and less than n/2. 11. An equalizing device, comprising: an internal equalizer equalizing the signal output from the fast Fourier transform unit based on the channel frequency response output from the channel estimation device; a decision feedback equalization compensator, which eliminates the intercarrier interference element of the signal output from the internal equalizer by a decision feedback equalization method; A linear minimum mean square error equalizer eliminates another intercarrier interference element of the signal output from the decision feedback equalization compensator by a linear minimum mean square error equalization method. 12. The equalizing apparatus of claim 11, wherein the internal equalizer comprises a divider to divide the signal output from the FFT unit by the channel frequency response output. 13. The equalization device of claim 11, wherein the decision feedback equalization compensator comprises: a decision device making a preliminary decision about the signal output from the internal equalizer; a multiplier for multiplying the signal output from the decision device with the channel slope vector output from the channel estimation device; A decision feedback equalization filter for filtering an intercarrier interference element produced by at least one subcarrier in the intercarrier interference element of the kth subcarrier of the signal output by the decision feedback equalization method from the multiplier; The subtractor subtracts the signal output from the decision feedback equalization filter from the signal output from the fast Fourier transform unit, k is a natural number greater than 1 and less than or equal to n, and n is the number of subcarriers. 14. The equalization device according to claim 13, wherein the linear minimum mean square error equalizer eliminates the intercarrier interference element of the kth subcarrier of the signal output from the decision feedback equalization compensator through a linear minimum mean square error equalization method Another intercarrier interference element generated by at least one other subcarrier. 15. The equalization device according to claim 14, wherein the linear minimum mean square error equalizer eliminates intercarrier interference elements generated by the kth subcarrier to the (k±L)th subcarrier, L is greater than or equal to 1 and A natural number less than n/2. 16. The equalization device according to claim 15, wherein the decision feedback equalization filter eliminates intercarrier interference elements generated by the kth subcarrier to the (k±Q)th subcarrier, Q is greater than or equal to 2 and less than n The natural number of /2. 17. The equalization apparatus of claim 16, wherein the decision feedback equalization filter removes intercarrier interference elements generated by (k±(L+1))th subcarriers to (k±Q)th subcarriers. 18. The equalizing apparatus of claim 11, wherein the linear minimum mean square error equalizer is an L-sublinear minimum mean square error equalizer. 19. The equalizing apparatus of claim 16, wherein L is a natural number smaller than Q. 20. An OFDM receiver comprising: The equalizing device as claimed in claim 11; The decoder receives and decodes the signal output from the equalization device.

Images